Ajay Pratap Yadav · Jul 23, 2026
A HELICS-based co-simulation framework enables seamless coordination among heterogeneous tools and models for integrating and testing electric drive models. The approach allows for reduced development time, flexible reuse of existing models, and efficient integration into software and Hardware-in-the-Loop environments. This scalable solution facilitates collaborative and repeatable electric drive system testing and development.
Why This Matters
This paper matters for power system engineers as it addresses interoperability challenges in electric machine systems, which is crucial for the effective design and testing of grid-scale systems, such as those involved in renewable integration and frequency regulation. The proposed HELICS-based co-simulation framework can facilitate more efficient and collaborative testing of electric drive models, ultimately contributing to a more resilient and adaptive power grid.
Sreejeet Maity, Aritra Mitra · Jul 23, 2026
Robust asynchronous Q-learning algorithms are proposed to defend against adversarial data corruption in reinforcement learning environments. The proposed {\texttt{BR-Async-Q}} algorithm uses batching and robust estimates of the Bellman optimality operator to achieve high-probability error bounds, matching those of vanilla Q-learning up to a small additive term. This provides the first robustness guarantee for asynchronous Q-learning subject to both reward and state corruption.
Why This Matters
This paper's focus on robust asynchronous Q-learning under reward and state corruption can enhance the reliability and stability of power grid operations, particularly in the context of renewable energy integration and cyber-physical attacks, by providing a more secure framework for optimizing control strategies and decision-making.
Shivanshu Tripathi, Maziar Raissi, Hamed Mohsenian-Rad · Jul 23, 2026
Continuously recorded high-resolution waveform measurements are used to identify events in power systems, which require automated methods for detection, localization, and classification. A new spectrogram-based framework is developed to solve this problem by transforming time-series waveforms into spectrograms, capturing transient and harmonic signatures more explicitly than raw data. The proposed method consistently improves event detection, localization, and classification over a baseline detector operating on raw measurements.
Why This Matters
This paper matters for power industry professionals as it presents a novel approach to automatically detecting and classifying events in high-resolution waveforms, which is crucial for grid operators to quickly respond to disturbances and faults, thereby improving the resilience and stability of the grid. This method can also inform utility planning and renewable integration strategies by enabling more accurate monitoring and analysis of power system dynamics.
Majid Dehghani, Arastoo H Salimi, Hamidreza Nazaripouya · Jul 22, 2026
A novel methodology for human-aware power restoration in distribution systems is proposed, incorporating probability of potential failures to guide resource allocation and ensuring a fair experience for customers. The approach balances restoration time across all failure locations through an adaptive partitioning policy, while also dispatching repair crews to accelerate service restoration with fairness as the goal. Simulation results show that this framework can deliver socially fair and customer-sensitive restoration outcomes in distribution networks under stochastic outage conditions.
Why This Matters
This paper matters for power industry professionals as it proposes a novel methodology for human-aware and fair service restoration in distribution systems, which can inform the development of more equitable and customer-centric outage management strategies, potentially improving public perception and satisfaction with grid operations, particularly in the context of renewable integration and capacity markets.
Cande Lian, Wentao Zeng, Jiabin Wu et al. · Jul 23, 2026
A new framework called FGDSE (Feature-Governed Dynamic Stacking Ensemble) is developed to predict electric vehicle charging infrastructure fault risk, enabling preventive maintenance in sustainable cities. The model surpasses 12 baselines in forecasting daily fault risk over 1 to 30 days and identifies extreme heat as a key climate stressor that amplifies its effect over time. FGDSE provides interpretable causal decision support with post-level treatment effects, helping strengthen urban mobility resilience and sustain low-carbon travel.
Why This Matters
This paper is relevant to power system engineers as it addresses the challenges of climate-resilient electric vehicle charging infrastructure, which directly impacts grid operations and resilience, particularly in terms of predictive maintenance and fault risk forecasting. The findings can inform utility planners' decisions on asset management and contribute to the development of more sustainable cities.
PV Magazine